Semantic Segmentation of Remote Sensing Images of Urban Areas using Deep Learning Methods

Abstract

As our world is attempting to step into a better innovative future, the need for smart cities is demanding.The major aspect of the smart city consists of environmental monitoring, innovative navigation, urban area planning and traffic management etc.Semantic segmentation of Remote Sensing images have a vital role in the development and the growth of smart cities. Remote sensing has been shown to be a very useful and efficient technique for mapping regions.In the pursuit of accurately detecting and segmenting specific regions on land, the collection of relevant data is as important as it sounds.Satellite imagery emerges as a paramount source, providing detectable insights for the detection and segmentation of areas.By harnessing the rich details in satellite images, precise information can be drawn.This approach allows us for a good analysis of the urban area.Given that the U-NET is currently a leading approach for accurate detection and segmentation, there is a growing trend to enhance its MIOU and overall performance by incorporating transformers into the model.This integration attempts to increase the model's precision and aid in its general improvement.